检索结果-内蒙古大学图书馆

Knowledge-embedding deep interpretable graph model for wear prediction: Application in pantograph-catenary systems

RELIABILITY ENGINEERING & SYSTEM SAFETY 2025年 261卷

作者： Mo, Yutao Peng, Yizhen Wu, Jing Fan, Kangbo Chongqing Univ Coll Mech & Vehicle Engn Chongqing 400044 Peoples R China Chongqing Rail Transit Grp Co LTD Chongqing 400042 Peoples R China

The wear modeling and life prediction of pantographs are crucial for ensuring the safety and reliability of urban rail transit systems. However, because of the complex interplay between stochastic vibrations and electrical currents, pantograph wear exhibits strong variability, and physics-based degradation predictions based solely on material parameters, environmental factors, and wear mechanisms are limited in accuracy. Purely data-driven approaches, on the other hand, are constrained by their reliance on large datasets and lack of interpretability, making them difficult to meet practical engineering needs. To address these challenges, we propose an interpretable variational model called IV-NBEATS. This study integrates the surface wear mechanism under the asperity hypothesis into the N-BEATS model using the projection principle, thereby enhancing the interpretability of the model. In addition, we introduce a method for describing the uncertainty of key wear parameters, enabling a deep network to represent the uncertainty of these parameters. Furthermore, to cope with dynamic changes in wear system parameters, we propose a dynamic updating method based on a Bayesian directed graph model that effectively overcomes the limitations of existing methods in capturing the temporal evolution of wear system parameters. Finally, the effectiveness of the proposed approach is demonstrated through the analysis of a real-world case study of pantograph wear.

关键词： Pantograph wear N -BEATS variational autoencoder Interpretability

来源：评论

学校读者我要写书评

暂无评论

DiffuseVAE plus plus : Mitigating training-sampling mismatch based on additional noise for higher fidelity image generation

引用

NEUROCOMPUTING 2025年 633卷

作者： Yang, Xiaobao Luo, Wei Ning, Hailong Zhang, Guorui Sun, Wei Ma, Sugang Xian Univ Posts & Telecommun Sch Comp Sci & Technol Xian Peoples R China Northwestern Polytech Univ Sch Comp Sci Xian Peoples R China

Denoising Diffusion Probabilistic Models (DDPMs) have demonstrated remarkable results in image generation. However, there exist a mismatch between the training and sampling process in current diffusion models, in addition, the U-Net denoising network based on simple residual blocks cannot predict noise information accurately, which affects the generated quality. To address these limitations, we present a novel image generation method that achieves higher fidelity. First, by additionally adding the standard Gaussian noise in the diffusion forward process, which does not disrupt the forward process, our method alleviates the mismatch. Subsequently, an important efficient denoising network based on U-Net is presented, where our proposed Simple Squeeze-Excitation and Simple GLU, combined with Depthwise Separable Convolution, enhance the ability of the model to predict real noise using the Simplified Nonlinear No Activation (SNNA) block. Furthermore, considering the structural characteristics of the baseline model, we introduce an additional cross-attention mechanism to enable DDPM to focus on VAE stage characteristics. Allowing the model to more accurately capture and learn the noise information. Finally, it is shown after extensive experiments the proposed DiffuseVAE++ obtains significant gains in FID scores, improving from 3.84 to 2.41 on CIFAR-10 and from 3.94 to 2.30 on CelebA-64. In particular, the IS scores on CIFAR-10 reaches 10.10, which is comparable to the current state-of-the-art methods competitively (e.g., U-ViT, StyleGAN2).

关键词： Image generation Generative model Diffusion model variational autoencoder Unconditional generation

来源：评论

学校读者我要写书评

暂无评论

Generative-based hybrid model with semantic representations for generalized zero-shot learning

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2025年第1期19卷 1-12页

作者： Akdemir, Emre Barisci, Necaattin Gazi Univ Grad Sch Informat Dept Comp Sci Ankara Turkiye Minist Natl Educ Directorate Gen Informat Technol Ankara Turkiye Gazi Univ Fac Technol Dept Comp Engn Ankara Turkiye

Generalized Zero-Shot Learning (GZSL) endeavors to recognize instances of seen and unseen classes using semantic information and labeled instances of only seen classes. In addressing the data imbalance problem in GZSL, generator networks produce synthetic features out of visual features and semantic features. However, the original semantic features exhibit insufficient distinctiveness. Throughout the current study, a hybrid generative-based GZSL framework that combines semantic descriptions used in fine-grained data sets with semantic attributes is proposed to increase the distinctiveness of the original semantic features. The proposed model offers semantic information enriched by the use of semantic attributes and semantic descriptions. A generalized classification is provided by using Generative Adversarial Network (GAN) and variational autoencoder (VAE) to generate visual features of unseen classes from this semantic information. The proposed method has been evaluated through extensive experiments on AWA2, CUB, FLO and SUN GZSL benchmark datasets. The experiments showed AWA2 70.96, CUB 76.71, FLO 92.12 and SUN 42.84% performance. Notably the proposed method achieved superior GZSL results, 3.41% and 3.52% on fine-grained CUB and FLO datasets respectively.

关键词： Generalized zero-shot learning Semantic representation variational autoencoder Generative adversarial network

来源：评论

学校读者我要写书评

暂无评论

Generating practical adversarial examples against learning-based network intrusion detection systems

引用

ANNALS OF TELECOMMUNICATIONS 2025年第3-4期80卷 209-226页

作者： Kumar, Vivek Kumar, Kamal Singh, Maheep Natl Inst Technol Dept Comp Sci & Engn Srinagar 246174 Uttarakhand India THDC Inst Hydropower Engn & Technol Dept Comp Sci & Engn Tehri 249124 Uttarakhand India IGDTUW Dept Informat Technol Delhi 110006 India

There has been a significant development in the design of intrusion detection systems (IDS) by using deep learning (DL)/machine learning (ML) methods for detecting threats in a computer network. Unfortunately, these DL/ML-based IDS are vulnerable to adversarial examples, wherein a malicious data sample can be slightly perturbed to cause a misclassification by an IDS while retaining its malicious properties. Unlike image recognition domain, the network domain has certain constraints known as domain constraints which are multifarious interrelationships and dependencies between features. To be considered as practical and realizable, an adversary must ensure that the adversarial examples comply with domain constraints. Recently, generative models like GANs and VAEs have been extensively used for generating adversarial examples against IDS. However, majority of these techniques generate adversarial examples which do not satisfy all domain constraints. Also, current generative methods lack explicit restrictions on the amount of perturbation which a malicious data sample undergoes during the crafting of adversarial examples, leading to the potential generation of invalid data samples. To address these limitations, a solution is presented in this work which utilize a variational autoencoder to generate adversarial examples that not only result in misclassification by an IDS, but also satisfy domain constraints. Instead of perturbing the data samples itself, the adversarial examples are crafted by perturbing the latent space representation of the data sample. It allows the generation of adversarial examples under limited perturbation. This research has explored the novel applications of generative networks for generating constraint satisfying adversarial examples. The experimental results support the claims with an attack success rate of 64.8%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy}

关键词： Adversarial example Domain constraints Deep learning Machine learning variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Uncertainty quantification for variational Bayesian dropout based deep bidirectional LSTM networks

引用

STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT 2025年 1-18页

作者： Sardar, Iqra Noor, Farzana Iqbal, Muhammad Javed Alsanad, Ahmed Akbar, Muhammad Azeem Int Islamic Univ Dept Math & Stat Islamabad 44000 Pakistan Univ Engn & Technol Taxila Dept Comp Sci Taxila 47050 Pakistan King Saud Univ Coll Comp & Informat Sci STCs Artificial Intelligence Chair Dept Informat Syst Riyadh 11451 Saudi Arabia Lappeenranta Lahti Univ Technol Software Engn Dept Lappeenranta 53851 Finland

Time series classification is a critical task in various domains, requiring robust models to handle inherent uncertainties in temporal data. These uncertainties, categorized as aleatoric and epistemic, pose significant challenges in achieving accurate predictions. In real-world applications, models often encounter unseen data that were not present during training process. Bayesian inference has been widely utilized for uncertainty quantification in statistics and machine learning. In this study, we proposed a Bayesian Deep Bi-LSTM model incorporating variational Bayesian dropout with a Gaussian prior and variational autoencoder (VAE). The proposed technique efficiently handles uncertainty in both the model and data while VAE reducing the dimensionality of model parameters. We apply this framework to univariate time series datasets from the UCR repository and compare its performance with four traditional machine learning methods and four sequential deep learning models. Experimental results demonstrate that the Bayesian deep Bi-LSTM model effectively improves overall classification performance. In particular, the model benefits significantly from data augmentation using SMOTE when handling imbalanced dataset. The variational Bayesian dropout model exhibits lower total uncertainty across both datasets, indicating more stable and reliable predictions compared to the VAE-based model. Future research should explore additional datasets from the UCR repository and investigate advanced uncertainty modeling techniques to further enhance performance and scalability.

关键词： Time series classification Bayesian deep Bi-LSTM Aleatoric and epistemic uncertainty variational Bayesian dropout variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Dynamical mode recognition of coupled flame oscillators by supervised and unsupervised learning approaches

引用

KNOWLEDGE-BASED SYSTEMS 2025年 307卷

作者： Xu, Weiming Yang, Tao Zhang, Peng City Univ Hong Kong Dept Mech Engn Kowloon Tong Kowloon Hong Kong 999077 Peoples R China Hong Kong Polytech Univ Dept Mech Engn Hung Hong Kowloon Hong Kong 999077 Peoples R China

Combustion instability in gas turbines and rocket engines, as one of the most challenging problems in combustion research, arises from the complex interactions among flames influenced by chemical reactions, heat and mass transfer, and acoustics. Identifying and understanding combustion instability is essential for ensuring the safe and reliable operation of many combustion systems, where exploring and classifying the dynamical behaviors of complex flame systems is a core task. To facilitate fundamental studies, the present work concerned dynamical mode recognition of coupled flame oscillators made of flickering buoyant diffusion flames, which have gained increasing attention in recent years but are not sufficiently understood. The time series data of flame oscillators were generated through fully validated reacting flow simulations. Due to the limitations of expertise-based models, a data-driven approach was adopted. In this study, a nonlinear dimensional reduction model of variational autoencoder (VAE) was used to project the high dimensional data onto a 2-dimensional latent space. Based on phase trajectories in the latent space, both supervised and unsupervised classifiers were proposed for datasets with and without well-known labeling, respectively. For labeled datasets, we established the Wassersteindistance-based classifier (WDC) for mode recognition;for unlabeled datasets, we developed a novel unsupervised classifier (GMM-DTW) combining dynamic time warping (DTW) and Gaussian mixture model (GMM). Through comparing with conventional approaches for dimensionality reduction and classification, the proposed supervised and unsupervised VAE-based approaches exhibit a prominent performance across seven assessment metrics for distinguishing dynamical modes, implying their potential extension to dynamical mode recognition in complex combustion problems.

关键词： variational autoencoder Dynamical mode recognition Coupled flames Wasserstein distance Dynamic time warping

来源：评论

学校读者我要写书评

暂无评论

A recombination generation method for missing fault data samples under unknown working conditions based on swap invariance-based multi-attribute disentangled representation network

引用

MEASUREMENT 2025年 249卷

作者： Cheng, Yujie Wang, Gaowei Gu, Haoxin Zhou, An Lu, Chen Ding, Yu Beihang Univ Sch Reliabil & Syst Engn Beijing 100191 Peoples R China Beihang Univ Hangzhou Int Innovat Inst Hangzhou Peoples R China Beihang Univ Sci & Technol Lab Reliabil & Environm Engn Beijing 100191 Peoples R China

Fault diagnosis under variable working conditions has become a prominent research focus in recent years. Numerous methods have been proposed to address it and data-driven methods have become popular. A datadriven diagnostic model with high accuracy is required to be trained on data that fully cover all working conditions. However, collecting such data is difficult, hindering the construction of diagnostic models under variable working conditions. Based on the theory that genetic traits are controlled by genes, this study proposes a recombination generation method for missing fault data samples under unknown working conditions based on a swap invariance-based multi-attribute disentangled representation network (SIMDRN). The proposed method can generate data samples under unknown working conditions based on the existing data under known working conditions. Firstly, three reconstruction losses and a generative adversarial loss are used to train the SIMDRN built using an improved variational autoencoder. Secondly, auxiliary samples are selected and encoded as disentangled factors using the trained SIMDRN. Finally, the missing samples under specific working conditions are generated by recombining and decoding the disentangled factors to improve the diagnostic performance of the models. The proposed method is verified using a superposed sine signal dataset and an experimental rollingbearing fault dataset. The results show that the samples generated by SIMDRN have great diversity and consistency. Moreover, the diagnostic accuracy is significantly improved by combining the generated samples with existing samples.

关键词： Fault diagnosis Variable working conditions variational autoencoder Disentangled representation learning Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

A generalized network with domain invariance and specificity representation for bearing remaining useful life prediction under unknown conditions

引用

KNOWLEDGE-BASED SYSTEMS 2025年 310卷

作者： Zheng, Qing Teng, Pengtao Zhang, Kai Ding, Guofu Lai, Xuwei Li, Zhixuan Yuan, Zhaocheng Southwest Jiaotong Univ Sch Mech Engn Chengdu 610031 Peoples R China Southwest Jiaotong Univ Technol & Equipment Rail Transit Operat & Maintena Chengdu 610031 Peoples R China Southwest Jiaotong Univ State Key Lab Rail Transit Vehicle Syst Chengdu 610031 Peoples R China Sichuan Special Antenna Mat & Adv Forming Engn Tec Mianyang 621000 Peoples R China Chengdu Inst Special Equipment Inspect & Testing Chengdu 610031 Peoples R China

The requirement to satisfy the assumption of identically and independently distributed data limits the practical application of traditional remaining useful life (RUL) prediction methods. In domain generalization methods, data augmentation and domain-invariant representations face the challenges of maintaining sequence consistency for long time-series data and prognosticating RUL under unknown operating conditions. A novel generalized network with invariance and specificity representation is proposed to address these issues. First, the runto-failure bearing data were divided into healthy and degraded stages. Next, the time-vector quantized variational autoencoder network generates diverse data from the degradation stage. Subsequently, the domain invariant and domain-specific feature representation network (DIDSR) is proposed for cross-condition prediction of rolling bearing RUL. The DIDSR network is divided into two parts: the domain-invariant component and the domain-specific component. The former extracts shared features from multiple source domains, while the latter learns the characteristics of each source domain through a domain classifier and weights them to obtain the final prediction. Ultimately, the proposed method is extensively evaluated on two bearing run-to-failure datasets. Comparative and ablation experiments validate the effectiveness and superiority of the approach.

关键词： Rotating machinery variational autoencoder Time series generation Domain-invariant representation learning Remaining useful life prognosis

来源：评论

学校读者我要写书评

暂无评论

Generalizable Prompts Guided by Image-Redundant Separation for Vehicle Reidentification

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第1期12卷 698-712页

作者： Kuang, Zhenyu Cheng, Lidong Zhang, Hongyang Huang, Yue Ding, Xinghao Xiamen Univ Sch Informat Minist Educ China Key Lab Multimedia Trusted Percept & Efficient Com Xiamen 361005 Peoples R China Xiamen Univ Inst Artificial Intelligence Minist Educ China Key Lab Multimedia Trusted Percept & Efficient Com Xiamen 361005 Peoples R China

Vehicle reidentification (reID) is a critical computer vision task with applications in video surveillance and autonomous vehicles. While significant progress has been made in recent years, domain generalization (DG) in reID remains a challenging and valuable research direction. Learning discriminative features that capture the intrinsic characteristics of vehicles, rather than domain-specific details, is paramount in addressing the domain shift problem, which encompasses disparities in data distribution, feature distribution, and label distribution. Recently, contrastive language image pretraining (CLIP) has attracted widespread attention because of its capacity to generalize knowledge across different domains or contexts. When fine-tuned for DG tasks, it can leverage this broad knowledge to perform well in domains or on tasks it has not specifically seen during training. The foremost work in this context is CLIP-reID, showcasing outstanding experimental performance on vehicle datasets through the integration of learnable prompts. However, the process of acquiring learnable prompts inevitably incorporates noisy text descriptions, such as background and camera style information, resulting in its limitations in DG tasks. To address this distinctive issue, we propose a CLIP-based Image-Redundant Separation (CIRS) framework to remove redundant domain-specific information and then implement visual-text alignment of CLIP. Specifically, we employ a classic variational autoencoder for image reconstruction, which can encourage the images generated by the vector quantized-variational autoencoder (VQ-VAE) network to contain features unrelated to vehicle IDs. Under the precise guidance of the image-redundant separation framework, a set of generalizable and learnable prompts for each vehicle can be effectively generated for reID. Extensive experimental results indicate that our method has achieved remarkable performance on several public datasets.

关键词： Context modeling Optimization Manuals Internet of Things Data models Contrastive language image pretraining (CLIP) domain generalization (DG) variational autoencoder vehicle reidentification (reID) Data models Contrastive language image pretraining (CLIP) domain generalization (DG) variational autoencoder vehicle reidentification (reID)

来源：评论

学校读者我要写书评

暂无评论

Abnormal behavior monitoring enhanced smart university stadium under the background of "Internet plus"

引用

INTERNET TECHNOLOGY LETTERS 2025年第3期8卷

作者： Li, Yan Meng, Xiao Zhang, Xiaochen Jilin Agr Sci & Technol Univ 77 Hanlin Rd Jilin Jilin Peoples R China Northeast Normal Univ Jilin Peoples R China

With the rapid development of the Internet of Things and 5G technology, smart university gymnasiums have become more and more important. However, it has become increasingly difficult for university gymnasium management, especially to detect abnormal behavior with dense crowds under limited venue space. To handle this issue, this paper designs an Artificial Intelligence Internet of Things (AIoT) abnormal behavior detection system which consists of the 5G camera, 5G transmission network and cloud platform. The 5G camera captures and transmits the video to the cloud platform by exploiting the 5G wireless sensor network. In the cloud platform, a hybrid variational autoencoder backbone which exploits the pre-trained VGG16 and Transformer model is deployed to detect abnormal behaviors. Moreover, by introducing adversarial training mechanisms, the robustness of the proposed model is effectively improved. The experimental results on our self-built gymnasium abnormal behavior dataset show that the proposed model can correctly identify most of the abnormal behaviors in the gymnasium compared to other models.

关键词： 5G technology abnormal behavior adversarial training internet plus technology smart university gymnasium variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：